Skip to main content
October 16, 2025

Parallel Task MCP Server

The Task MCP Server uses a first-of-its-kind async architecture that lets agents start research tasks and continue executing other work without blocking. This is critical for production agents handling complex workflows— start a deep research task on competitor analysis, move on to enriching a prospect list, then retrieve the research results when complete.
The Task MCP Server can be useful for professionals who want to bring the power of Parallel’s Tasks to their preferred MCP client, or for developers who are building with Parallel Tasks.Learn more in the release blog.
October 12, 2025

Core2x Processor

The new Core2x processor is now available for the task API. Core2x bridges the gap between Core and Pro processors for better cost control on Task runs.
Use Core2x for:
  • Cross-validation across multiple sources without deep research level exploration
  • Moderately complex synthesis where Core might fall short
  • Structured outputs with 10 fields requiring verification
  • Production workflows where Pro’s compute budget exceeds requirements
Learn more in the release blog.
October 6, 2025

Enhanced Basis Features Across All Processors

All Task API processors now provide complete basis verification with Citations, Reasoning, Confidence scores, and Excerpts. Previously, lite and base processors only included Citations and Reasoning, while core and higher tiers provided the full feature set. This enhancement enables comprehensive verification and transparency across all processor tiers, making it easier to validate research quality regardless of which processor you choose.
With this update, even the most cost-effective lite processor now returns:
  • Citations: Web URLs linking to source materials
  • Reasoning: Detailed explanations for each output field
  • Confidence: Calibrated reliability ratings (high/medium/low)
  • Excerpts: Relevant text snippets from citation sources
This improvement supports more effective hybrid AI/human review workflows at every price point.Learn more in the release blog.
September 16, 2025

TypeScript SDK

The Parallel TypeScript SDK is now generally available for the Task and Search API - providing complete type definitions, built in retries, timeouts, and error handling, and custom fetch client support. Learn more in our latest blog.
September 11, 2025

Deep Research Reports

Parallel Tasks now support comprehensive markdown Deep Research report generation. Every Deep Research report generated by Parallel comes with in-line citations and relevant source excerpts for full verifiability. Simply enable output_schema: text to get started. Learn more in our latest blog.
September 9, 2025

Expanded Deep Research Benchmarks

Today we are releasing expanded results that demonstrate the complete price-performance advantage of Parallel Deep Research - delivering the highest accuracy across every price point.On Browsecomp:
  • Parallel Ultra achieves 45% accuracy at up to 17X lower cost
  • Ultra8x achieves state-of-the-art results at 58% accuracy
On DeepResearch Bench:
  • Parallel Ultra achieves an 82% win rate against reference compared to GPT-5 at 66%, while being half the cost
  • Ultra8x achieves a 96% win rate
Learn more in our latest blog.
August 21, 2025

Webhooks for Tasks

Webhooks are now available for Parallel Tasks. When you’re orchestrating hundreds or thousands of long-running web research tasks, webhooks push real-time notifications to your endpoint as tasks complete. This eliminates the need for constant polling. Learn more in our latest blog.
August 14, 2025

Deep Research Benchmarks

Today, we’re announcing that Parallel is the only AI system to outperform both humans and leading AI models like GPT-5 on the most rigorous benchmarks for deep web research. Our APIs are now broadly available, bringing production-grade web intelligence to any AI agent, application, or workflow. Learn more in our latest blog.
August 7, 2025

Server-Sent Events for Tasks

Server-Sent Events are now available for Parallel Task API runs. SSE delivers live progress updates, model reasoning, and status changes as tasks execute. Learn more in our latest blog.
August 5, 2025

New advanced deep research processors

New advanced processors are now available with Parallel Tasks, giving you granular control over compute allocation for critical research workflows. Last month, we demonstrated that accuracy scales consistently with compute budget on BrowseComp, achieving 39% and 48% accuracy with 2x and 4x compute respectively. These processors are now available as ultra2x and ultra4x, alongside our most advanced processor yet - ultra8x. Learn more in our latest blog.
August 4, 2025

Auto Mode in Parallel Tasks

Parallel Tasks now support Auto Mode, enabling one-off web research queries without requiring explicit output schemas. Simply ask a question. Our processors will then conduct research and generate a structured output schema for you. Learn more in our latest blog.
July 31, 2025

State-of-the-art Search API benchmarks

The Parallel WebTools MCP Server, built on the same infrastructure as the Parallel Search API, demonstrates superior performance on the WISER-Search benchmark while being up to 50% cheaper. Learn more in our latest blog.

Parallel WebTools MCP Server in Devin

The Parallel WebTools MCP Server is now live in Devin’s MCP Marketplace, bringing high quality web research capabilities directly to the AI software engineer. With a web-aware Devin, you can ask Devin to search online forums to debug code, linear from online codebases, and research APIs. Learn more in our latest blog.
July 28, 2025

Tool Calling via MCP Servers

Parallel Tasks now support Tool Calling via MCP Servers. With a single API call, you can choose to expose tools hosted on external MCP-compatible servers and invoke them through the Task API. This allows Parallel agents to reach out to private databases, code execution sandboxes, or proprietary APIs - without custom orchestrators or standalone MCP clients. Learn more in our latest blog.
July 14, 2025

The Parallel WebTools MCP Server

The Parallel WebTools MCP Server is now generally available, making our Search API instantly accessible to any MCP-aware model as a drop-in tool. This hosted endpoint takes flexible natural language objectives as inputs and provides AI-native search results with extended webpage excerpts. Built on Parallel’s proprietary web infrastructure, it offers plug-and-play compatibility with OpenAI, Anthropic, and other MCP clients at production scale. Learn More.
July 8, 2025

Source Policy for Task API and Search API

Source Policy is now available for both Parallel Tasks and Search API - giving you granular control over which sources your AI agents access and how results are prioritized. Source Policy lets you define exactly which domains your research should include or exclude. Learn more in our latest blog.
July 2, 2025

Task Group API in beta

Today we’re launching the Task Group API in public beta for large-scale web research workloads. When your pipeline needs hundreds or thousands of independent Parallel Tasks, the new Group API wraps operations into a single batch with unified monitoring, intelligent failure handling, and real-time results streaming. These batch operations are ideal for bulk CRM enrichment, due diligence, or competitive intelligence workflows. Learn more in our latest blog.
June 17, 2025

State of the Art Deep Research APIs

Parallel Task API processors achieve state-of-the-art performance on BrowseComp, a challenging benchmark built by OpenAI to test web search agents’ deep research capabilities. Our best processor (ultra) reaches 27% accuracy, outperforming human experts and all commercially available web search and deep research APIs - while being significantly cheaper. Learn more in our latest blog.
August 05, 2025

Search API in beta

The Parallel Search API is now available in beta - providing a tool for AI agents to search, rank, and extract information from the public web. Built on Parallel’s custom web crawler and index, the Search API takes flexible inputs (search objective and/or search queries) and returns LLM-ready ranked URLs with extended webpage excerpts. Learn more in our latest blog.
curl https://api.parallel.ai/v1beta/search \
  -H "Content-Type: application/json" \
  -H "x-api-key: ${PARALLEL_API_KEY}" \
  -d '{
    "objective": "When was the United Nations established? Prefer UN'\''s websites.",
    "search_queries": [
      "Founding year UN",
      "Year of founding United Nations"
    ],
    "processor": "base",
    "max_results": 5,
    "max_chars_per_result": 1500
  }'

  • [Platform] Fixed an issue where copy paste URL actions were incorrectly identified as copy paste CSV actions.
May 30, 2025

Chat API in beta

Parallel Chat is now generally available in beta. The Chat API utilizes our rapidly growing web index to bring real-time low latency web research to interactive AI applications. It returns OpenAI ChatCompletions compatible streaming text and JSON outputs, and easily drops in to new and existing web research workflows. Learn more in our latest blog.
from openai import OpenAI

client = OpenAI(
    api_key="PARALLEL_API_KEY",  # Your Parallel API key
    base_url="https://api.parallel.ai"  # Parallel's API beta endpoint
)

response = client.chat.completions.create(
    model="speed", # Parallel model name
    messages=[
        {"role": "user", "content": "What does Parallel Web Systems do?"}
    ],
    response_format={
        "type": "json_schema",
        "json_schema": {
            "name": "reasoning_schema",
            "schema": {
                "type": "object",
                "properties": {
                    "reasoning": {
                        "type": "string",
                        "description": "Think step by step to arrive at the answer",
                    },
                    "answer": {
                        "type": "string",
                        "description": "The direct answer to the question",
                    },
                    "citations": {
                        "type": "array",
                        "items": {"type": "string"},
                        "description": "Sources cited to support the answer",
                    },
                },
            },
        },
    },
)

print(response.choices[0].message.content)
  • [Task API] Fixed an issue where the Task API was returning malformed schema formats.
May 21, 2025

Basis with Calibrated Confidences

Basis is a comprehensive suite of verification tools for understanding and validating Task API outputs through four core components.
  1. Citations: Web URLs linking directly to source materials.
  2. Reasoning: Detailed explanations justifying each output field.
  3. Excerpts: Relevant text snippets from citation URLs.
  4. Confidences: A calibrated measure of confidence classified into low, medium, or high categories.
Use Basis with Calibrated Confidences to power hybrid AI/human review workflows focused on low confidence outputs - significantly increasing leverage, accuracy, and time efficiency. Read more in our latest blog post.
{
 "field": "revenue",
 "citations": [
   {
     "url": "https://www.microsoft.com/en-us/Investor/earnings/FY-2023-Q4/press-release-webcast",
     "excerpts": ["Microsoft reported fiscal year 2023 revenue of $211.9 billion, an increase of 7% compared to the previous fiscal year."]
   },
   {
     "url": "https://www.sec.gov/Archives/edgar/data/789019/000095017023014837/msft-20230630.htm",
     "excerpts": ["Revenue was $211.9 billion for fiscal year 2023, up 7% compared to $198.3 billion for fiscal year 2022."]
   }
 ],
 "reasoning": "The revenue figure is consistent across both the company's investor relations page and their official SEC filing. Both sources explicitly state the fiscal year 2023 revenue as $211.9 billion, representing a 7% increase over the previous year.",
 "confidence": "high"
}

Billing Upgrades

We’ve made several improvements to help you more seamlessly manage and monitor Billing. This includes:
  • Auto-reload: Avoid service interruptions by automatically adding to your balance when configured thresholds are met.
  • Billing History: View prior Invoices and Receipts. Track status, amount charged, and timestamp of charges.

  • [Task API] Top-level output fields now correctly return null when appropriate, rather than lower-level fields returning empty string.
  • [Task API] Improved pro and ultra responses for length list-style responses.
  • [Platform] The improved Parallel playground is now available by default at platform.parallel.ai/play instead of platform.parallel.ai/playground.
April 24, 2025

Task API for web research

Parallel Tasks enables state-of-the-art web research at scale, with the highest quality at every price point. State your research task in natural language and Parallel will do the rest of the heavy lifting - generating input/output schemas, finding relevant URLs, extracting data in a structured format.
from parallel import Parallel
from pydantic import BaseModel, Field

class ProductInfo(BaseModel):
    use_cases: str = Field(
        description="A few use cases for the product."
    )
    differentiators: str = Field(
        description="3 unique differentiators for the product as a bullet list."
    )
    benchmarks: str = Field(
        description="Detailed benchmarks of the product reported by the company."
    )

client = Parallel()
result = client.task_run.execute(
    input="Parallel Web Systems Task API",
    output=ProductInfo,
    processor="core"
)

print(f"Product info: {result.output.parsed.model_dump_json(indent=2)}\n")
print(f"Basis: {'\n'.join([b.model_dump_json(indent=2) for b in result.output.basis])}")

Python SDK

Our SDK is now available for Python, making it easy to implement Parallel into your applications. The Python SDK is at parity with our Task API endpoints and simplifies request construction and response parsing.

Flexible Processors

When running Tasks with Parallel, choose between 5 processors - lite, base, core, pro, and ultra. We’ve built distinct processor options so that you can optimize price, latency, and quality per task.

Self-Serve Developer Platform

Platform is the home for Playground, API Keys, Docs, Billing, Usage, and more.
  • Run a research task from scratch or using a template from Task Library
  • Generate and manage API keys for secure integration
  • Manage billing details, auto-reload settings, and usage analytics
  • Access comprehensive guides to learn how to use the API